CVI/example_usage.py at main · ppIntelligence/CVI · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
import os
import torch
import numpy as np
from pathlib import Path
import sys

project_root = Path(__file__).parent.parent
sys.path.append(str(project_root))

from cvi.training.trainer import CVITrainer
from cvi.data.dataset import DataModule
from cvi.utils.config import load_config
from cvi.utils.utils import set_seed, get_device


def create_sample_data(data_dir: str, num_samples: int = 10):
    import cv2
    from PIL import Image

    os.makedirs(os.path.join(data_dir, 'train', 'images'), exist_ok=True)
    os.makedirs(os.path.join(data_dir, 'train', 'masks'), exist_ok=True)
    os.makedirs(os.path.join(data_dir, 'test', 'images'), exist_ok=True)
    os.makedirs(os.path.join(data_dir, 'test', 'masks'), exist_ok=True)

    for i in range(num_samples):
        img = np.random.rand(512, 512) * 255
        img = img.astype(np.uint8)

        mask = np.zeros((512, 512), dtype=np.uint8)

        cv2.line(mask, (50, 50), (450, 450), 255, 3)
        cv2.line(mask, (100, 400), (400, 100), 255, 2)
        cv2.circle(mask, (256, 256), 50, 255, 2)

        noise = np.random.rand(512, 512) * 50
        mask = np.clip(mask.astype(np.float32) + noise, 0, 255).astype(np.uint8)

        cv2.imwrite(os.path.join(data_dir, 'train', 'images', f'image_{i:03d}.png'), img)
        cv2.imwrite(os.path.join(data_dir, 'train', 'masks', f'mask_{i:03d}.png'), mask)

        test_img = np.random.rand(512, 512) * 255
        test_img = test_img.astype(np.uint8)

        test_mask = np.zeros((512, 512), dtype=np.uint8)
        cv2.line(test_mask, (100, 100), (400, 400), 255, 2)
        cv2.rectangle(test_mask, (200, 200), (300, 300), 255, 2)

        cv2.imwrite(os.path.join(data_dir, 'test', 'images', f'test_{i:03d}.png'), test_img)
        cv2.imwrite(os.path.join(data_dir, 'test', 'masks', f'test_mask_{i:03d}.png'), test_mask)


def main():
    print("CVI Framework Example")
    print("=" * 50)

    set_seed(42)

    device = get_device()
    print(f"Using device: {device}")

    data_dir = "./sample_data"
    print(f"Creating sample data in {data_dir}...")
    create_sample_data(data_dir, num_samples=5)

    config_path = "cvi/configs/default.yaml"
    print(f"Loading configuration from {config_path}...")
    config = load_config(config_path)

    config['dataset']['data_root'] = data_dir
    config['dataset']['train_images'] = 'train/images'
    config['dataset']['train_masks'] = 'train/masks'
    config['dataset']['test_images'] = 'test/images'
    config['dataset']['test_masks'] = 'test/masks'
    config['device'] = str(device)

    config['training']['num_epochs_per_phase'] = 2
    config['training']['batch_size'] = 1
    config['num_synthetic_samples'] = 10

    print("Configuration loaded successfully!")
    print(f"Model: {config['model_id']}")
    print(f"Device: {config['device']}")
    print(f"Training epochs per phase: {config['training']['num_epochs_per_phase']}")

    print("\nInitializing CVI trainer...")
    trainer = CVITrainer(
        config=config,
        device=str(device),
        use_wandb=False,
        log_dir="./example_logs"
    )

    print("Creating data module...")
    data_module = DataModule(config)

    datasets = data_module.create_datasets()
    print(f"Created {len(datasets)} datasets: {list(datasets.keys())}")

    data_loaders = data_module.create_data_loaders(datasets)
    print(f"Created {len(data_loaders)} data loaders")

    test_loader = data_module.create_test_loader()
    print(f"Test loader created with {len(test_loader)} samples")

    print("\nStarting training...")
    try:
        results = trainer.train_complete(
            train_loaders=data_loaders,
            val_loader=test_loader,
            num_epochs_per_phase=config['training']['num_epochs_per_phase'],
            save_interval=1
        )

        print("Training completed successfully!")
        print(f"Results: {list(results.keys())}")

        print("\nEvaluating model...")
        final_metrics = trainer.evaluate(test_loader)
        print("Final evaluation metrics:")
        for metric, value in final_metrics.items():
            print(f"  {metric}: {value:.4f}")

    except Exception as e:
        print(f"Training failed with error: {e}")
        print("This is expected for the demo with limited data and epochs")

    trainer.close()

if __name__ == '__main__':
    main()